Industrial IoT

New ‘Event Explorer’ helps uncover anomalies in event data

3 min readMost connected product manufacturers and service providers realize very early in their Internet [...]

3 min read

Most connected product manufacturers and service providers realize very early in their Internet of Things (IoT) journey that they will be gathering a lot of data, and that they will need to eventually ‘make sense’ of this data to improve the services they provide and the products they sell. But the process of identifying patterns, anomalies and trends from data is never clear-cut. It can be best described using three ‘dimensions’ or phases:

  1. The first dimension consists of reviewing what you know. During this assessment, you will compare the current understanding of your ecosystem of products with the new information you acquired throughout the products connected life. This will confirm (or contradict) the validity of the perceptions and/or facts you formed earlier.
  2. In the second dimension, you will consider what you know you don’t know. You will navigate through product data using predefined questions and assumptions. The ultimate goal here is to improve product development, enhance existing products with new features, identify upsell opportunities, decrease truck rolls, reduce downtime, set up usage-based pricing strategies, etc.
  3. The third dimension is where you encounter what you don’t know you don’t know. This is the phase where you discover you probably don’t know as much as you think you do. In this phase, focus is put on trying to identify new business trends and needs, creating outcome-based services (e.g. comfort as a service, irrigation as a service), etc.

Whereas mnubo SmartObjects already includes plugins to many open source and commercial artificial intelligence & machine learning tools (e.g. Python, R, Keras, etc.) to help explore and make sense of IoT data, these involve coding and, rightfully, a data science background. While very useful for some use cases, we also believe that non-coders should be able to quickly and easily explore their IoT data and answer questions like ‘my products are connected and generating data, now what? With this objective in mind, SmartObjects’ new Event Explorer includes user-friendly graphical tools to:

    1. Analyse automatically an IoT data set in its entirety, or any sub set, and highlight the distribution of all events and the correlation between one another; events managed by SmartObjects include, amongst others:
      • Events from the connected products themselves (e.g. temperature measured, diagnostic report sent, part replaced, etc.)
      • Events from mobile devices remotely controlling the products and from other surrounding products like control panels, gateways, etc. (e.g. temperature increased, fan stopped, mobile device paired, profile updated, etc.)
      • Events from IoT device management and application solutions (e.g. firmware updated, diagnostic requested, object created, orchestration rule triggered, etc.)
      • Events triggered by humans, like operators, service technicians or customers (e.g. door opened, button pressed, filter replaced, etc.)
      • AI-generated events (e.g. from smart voice-controlled assistants, automatic rules, etc.)


    1. Cluster events by any time period to highlight discrepancies, anomalies or trends


  1. Create collections of products and customers that are similar or dissimilar with regards to their behaviour, usage, performance, etc.

Key business use cases enabled by this new Event Explorer include:

  • Enabling a comprehensive view on product usage and performance
  • Identifying gaps in customer experience that tarnish the overall experience
  • Identifying undisclosed downtimes that negatively influence the overall experience and operating costs
  • Quickly identifying profiles of customers likely to churn, or ripe for an upsell
  • Establishing if products are used as they were meant to be used
  • Leveraging product usage feedback to focus R&D spend and enhance existing products with new features
  • Identifying unpopular hardware/software versions, etc.

If you are interested in getting more information about other industrial IoT insights, check out this link